Design of a Drone with a Robotic End-Effector

Suhas Varadaramanujan, Sawan Sreenivasa, Praveen Pasupathy, Sukeerth Calastawad,

Melissa Morris, Sabri Tosunoglu

Department of Mechanical and Materials Engineering

Florida International University
Miami, Florida 33174

svara008@fiu.edu, ssree005@fiu.edu , ppasu001@fiu.edu, scala007@fiu.edu,

mmorr009@fiu.edu, tosun@fiu.edu

ABSTRACT example, the widely-used predator drone for military purposes is

The concept presented involves a combination of a quadcopter
drone and an end-effector arm, which is designed with the
capability of lifting and picking fruits from an elevated position.
The inspiration for this concept was obtained from the swarm
robots which have an effector arm to pick small cubes, cans to even
collecting experimental samples as in case of space exploration.
The system as per preliminary analysis would contain two
physically separate components, but linked with a common
algorithm which includes controlling of the drone’s positions along
with the movement of the arm.
Keywords
Drone, end effector, autonomous, payload.
1. INTRODUCTION
A robot is a machine that is programmable and capable of carrying
out a series of complex actions automatically. Robots can be guided
by an external control device or the control may be embedded
within. Robots may be constructed to take on a human form, but
most robots are machines designed to perform a task with little
regard to how they look. Robots can be autonomous or semi-
autonomous such as Honda’s Advanced Step in Innovative
Mobility, also known as ASIMO, and TOSY’s Ping Pong Playing
Robot (TOPIO) which are marvels in the field of robotics. In
addition to this there is considerable progress seen in area of
industrial robots, medical robots, patent assisting robots, human
rehabilitation, athletic training robots, and a whole set of more
applications. In addition to land-based robots there is also
developments in the bots used for aerospace and marine
applications. By mimicking a lifelike appearance or automating
movements, a robot may convey a sense of intelligence or thought
of its own.
An unmanned aerial vehicle (UAV), commonly known as a drone,
is an aircraft without a human pilot aboard. UAVs are a component
of an unmanned aircraft system (UAS) which include a UAV, a
ground-based controller, and a system of communications between
the two. The flight of UAVs may operate with various degrees of
autonomy either under remote control by a human operator, or fully
or intermittently autonomously by on-board computers. For
the MQ-1 by General Atomics which is remote controlled, UAVs
typically fall into one of six functional categories (i.e. target and
decoy, reconnaissance, combat, logistics, R&D, civil and
commercial).
With the advent of aerial robotics technology, UAVs became more
sophisticated and led to development of quadcopters which gained
popularity as mini-helicopters. A quadcopter, also known as a
quadrotor helicopter, is lifted by means of four rotors. In operation,
the quadcopters generally use two pairs of identical fixed pitched
propellers; two clockwise (CW) and two counterclockwise (CCW).
They use independent variation of the speed of each rotor to achieve
control. By varying the speed of each rotor, it is possible to
specifically generate a desired total thrust, to locate for the center
of thrust both laterally and longitudinally, and to create a desire
total torque or turning force. In addition to this development,
quadcopters were designed to adopt an end effector. In robotics, an
end effector is the device at the end of a robotic arm designed to
interact with the environment. The exact nature of this device
depends on the application of the robot.
In the strictest definition, which originates from serial robotic
manipulators, the end effector is the last link (end) of the robot. At
this endpoint, the tools are attached. In a wider sense, an end
effector is the part of a robot that interacts with the work
environment. This does not refer to the wheels of a mobile robot
nor the feet of a humanoid robot which are not end-effectors
because they are part of the robot's mobility.
End effectors may consist of a gripper or a tool. When referring to
robotic end-effectors, there are four broad categories of robot
grippers. These are (a) Impactive – jaws or claws which physically
grasp by direct impact upon the object, (b) Ingressive – pins,
needles or hackles which physically penetrate the surface of the
object (used in textile, carbon and glass fiber handling), (c)
Astrictive – suction forces applied to the objects surface (whether
by vacuum, magneto- or electro adhesion) and (d) Contigutive –
requiring direct contact for adhesion to take place (such as glue,
surface tension or freezing) [10].

30th Florida Conference on Recent Advances in Robotics, May 11-12, 2017, Florida Atlantic University, Boca Raton, Florida
2. LITERATURE SURVEY
Mendoza et al. [1] introduced a conceptual design which involved
using multiple quadrotors programmed to make use of the yaw
movement, linked to each other using a bar containing end effectors
as seen in Figure 1. This paper aimed to introduce and spread an
extensive mathematical formality that is used to build a new kind
of UAV manipulator. The main contribution of this paper was to
introduce the possibility of linkages that could be created between
two systems.
Figure 1. Air-arm and end-effector
A parallel manipulator with robustly maintained precise end
effector position stowed below a quadrotor UAV used for lifting of
light weight mass was developed by Danko et al. [2] They also
proposed, constructing a six degree of freedom parallel manipulator
which is used to robustly maintain precise end effector positions
despite post-UAV perturbations. The parallel manipulator allows
for very little moving mass and is easily stored below a quad rotor
UAV as can be visualized from Figures 2 and 3.
Figure 2. Parallel manipulator extended
Figure 3. Parallel manipulator stowed
30th Florida Conference on Recent Advances in Robotics, May 11-12, 2017, Florida Atlantic University, Boca Raton, Florida
Paul et al. [3] introduced the concept of a hyper-redundant
manipulator as seen in Figure 4 to be used for the mobile
manipulating UAVs. The flexibility of the links was observed by
authors when the arm was programmed in MATLAB for moving
in the shape of eight as seen in Figure 5 for testing the controllers.
Figure 4. Hyper redundant arm
It can be controlled in such a way that links is moved within the
arm’s free space to help reduce negative impacts on the host
platform’s stability and the end effector to track environmental
objects smoothly despite host platform motions. The hyper
redundant manipulator does not include a spherical wrist, even
though through preliminary reachability analysis, the manipulator
had an improved reachable arm volume by over 20 %.
Figure 5. Hyper-redundant arm sweeping in the shape of “8”
Maier et al. [4] discussed about an approach allowing independent
control of position and orientation of a UAV, whereby an arbitrary
stable attitude controller could be used. The main contribution of
this literature was to propose a new control approach for the entire
system composed of Vertical Take and Landing (VTOL) UAV and
a manipulator. The advantage of the presented approach is that the
interaction forces between robots and UAV were considered
explicitly and that a Lyapunov stability proof for the UAV
subsystem could be derived directly. This approach extends the
classical Cartesian impedance controller to account for the UAV’s
rigid body dynamics.
Zisimatos et al. [5] proposed a robotic hand which is efficient in
grasping a series of everyday life objects, is general purpose, and is
validated using series of experimental paradigms. This approach
further improves the design of conventional grippers that are
commonly used for grasping, both in industry and research due to
its low complexity and relatively low cost. The authors introduced
a new end effector system capable of lifting heavy weights and with
grippers varying from two fingers to four fingers with a weight of
0.088 lb. to 0.53 lb. respectively as seen in Figure 6.

Figure 7. Fruit picking robot

Sarig [8] discusses the technological improvements over the past

decade related to fruit picking and the way robots have been
programmed to locate the fruit in specified area without any
damage to the fruit. The author developed an end effector with six
degrees of freedom, having three rotational joints connecting
neighboring links to cover the defined workspace. The author has
made use of the Puma 560 robot arm which has forward and reverse
kinematics whose controls are programmed through MATLAB
Robotic Tool Box. The joint angles, velocities and torques of the
robot arm were studied during example pick cycles which span the
work space of representative peach and orange trees.

Figure 6. Different robot hands created using identical

modular fingers and the modular fingers basis
Brown et al. [6] studied the kinematic and dynamic behavior of
humans lifting heavy weights and applied it to a robotic arm which
was made to lift heavy weights as a part of dynamic task. The
authors discussed the kinematic and dynamic behavior of the robot
with experimental results.
In addition to fruit picking robot as specified previously further
developments was done by Henten et al. [9] who discussed the
specific application of cucumber picking robots and the efficiency
of the end effector. The manipulator was designed with seven
degrees of freedom. The robot is equipped with computer vision
system which could detect more than 95% of cucumbers in a green
house. The ripeness of the cucumbers was determined based on the
geometric models.

The results showed that the arm could lift 20 kg mass with the links
having kinematics like that of human muscle. This approach uses a
new methodology that is dynamic motion control. This technology
is applicable not only for lifting but for pulling, peeling and
destroying tasks.

Bhope et al. [7] gives an overall review of robot application

employed over a period. The paper also provides information on
the type of end effector used for fruit picking with pictorial
representation as shown in Figure 7.

The robot mentioned in this paper deals with applications such as Figure 8. Vegetable harvest robot in the greenhouse
spraying and motorized weed control, fruit selection and inspecting
the farms day and night for an efficient result which in turn reduces
the farmer’s effort. One of the advantages of the smaller machines 3. OBJECTIVES OF PROPOSED WORK
employed in agriculture is that they are more satisfactory to the
non-farm population. The objectives of the proposed drone model are:
1. To help the physically handicapped person pick an object of
necessity (Water bottle, food items, fruits etc.) from a far-off
distance and bring it to his/her vicinity.

30th Florida Conference on Recent Advances in Robotics, May 11-12, 2017, Florida Atlantic University, Boca Raton, Florida
2. To help farmers in harvesting the fruits and vegetables thereby costlier than the conventional method wherein the arms are built
reducing the distance of travel in long fields and on a minor using traditional manufacturing methods which is even more time
scale help develop automated farming. consuming but less costly when compared to former. It is important
3. Along with this the drone can be used to spray fertilizers and to note that most of the products that will be used for prototyping
pesticides over the crops, thereby reducing health hazards of the proposed model are 3D printed.
farmers. 4.1. Prototype Modeling
4. MODEL SPECIFICATION
Based on the literature survey conducted, it was found that
developing the prototype to suit the purpose of agriculture and
domestic application is more effective in this situation as it would
reduce the complexity of modeling and would be a significant
innovation in the field of automated farming. Modeling of a drone
for applications such as military and industrial would be complex.
Drone Parameters
1. Number of rotors: 4
2. Wings span: 23.22 in
3. Weight: 1.890 Kg (2.84 lbs.)
4. Electronic Speed controller: 20A
5. Arm material: Acrylonitrile Butadiene Styrene BS – A
(ABS) Figure 9. Isometric view of the drone
6. Propeller Material: ABS
7. Payload Capacity: 450 g (0.99 lbs.)
8. Length of arm: 11.81 in
9. Effector radius: 2.36 in

Camera Parameters
1. Height in pixels: 3000
2. Width in pixels: 4000
3. Lens focal length in mm: 3.61
4. Lens field of view: 940

Controller
1. Control System: Remote
2. Operating Range: 2000m
3. Controller Frequency: 2.4Hz
4. Controller Voltage: 7.4V

Battery Figure 10. Right view of the drone

1. Battery Capacity: 7000 mAh
2. Battery type: Lithium Polymer 4S
3. Battery Power: 14.8V

Figure 9 represents the isometric view of the drone that has been
modeled considering the specifications as given above. A clear
picture of the placement of components has been shown in Figure
10, which shows that the drone has been designed carefully with
each component being reserved with spaces meant for them. Figure
11 shows the assembled view of the drone with the end effector
arm. The arm is enabled with a 2- DOF link with the effector having
a rotational DOF which is equipped with suction cups to hold the
payload firmly. Figures 12-17 represent the stress analyses that was
performed when the drone was made to carry loads varying in mass
and major concentration in the arms. Both displacement and Von-
Mises simulation results have been presented with figures
indicating that as the load increases, the displacement in the drone’s
arm increases and hence leading to the conclusion that a restricted
amount of load can be carried. The arm could be manufactured
using 3D printing technology and could be time-consuming and Figure 11. Assembled view of the drone

30th Florida Conference on Recent Advances in Robotics, May 11-12, 2017, Florida Atlantic University, Boca Raton, Florida
4.2.1 Displacement simulation results
The figures shown below indicate the displacement or the drone
arm with increase in the mass of the payload lifted. The payload
masses range from 0.75 N to 3N with intervals of 0.75 N.

Figure 14. Displacement results for 2.25 N

Figure 12. Displace results for 0.75 N

Figure 15. Displacement results for 3 N

Figure 14 and 15 show the displacement for 0.50 lbs. and 0.67 lbs.
Figure 13. Displacement for 1.5 N for a factor of safety of 2.5.

Figure 12 and 13 show the displacement for 0.1686 lbs. and 0.3372 4.2.2 Von-Mises simulation results
lbs. respectively. Figure 12 shows that the load factor near the
This section illustrates the results of Von-Mises stresses that occur
propeller region is critical. As the motor provides downward thrust,
in the drone arm which increase with the increase in load that is
the beam must withstand the thrust near the propeller. When the
carried.
analysis was carried out for 0.1686 lbs., the value of displacement
obtained was about 9.645*10-5 m. The factor of safety of the ABS
material is about 2.5. Similarly, for 0.3372 lbs. the displacement
obtained was about 0.001929 m.

30th Florida Conference on Recent Advances in Robotics, May 11-12, 2017, Florida Atlantic University, Boca Raton, Florida
 Figure 17. Von-Mises results for 3 N
Figure 15. Von-Mises results for 0.75 N
Figures 16 and 17 show the stress analysis for varying loads. From
the analyses conducted, it was concluded that the Von-Mises
stresses increase with increase in the load applied on the arm. The
arm is being designed to carry light weight loads and the results of
analyses show that loading the arm with furthermore load results in
a structural damage or failure. For higher loads, we obtained
1.8636*106 Pa which was an ideal value for our proposed design.

5. COST OF MANUFACTURING
The estimated costs of each parts that are used to build the drone
model are shown in Table 1. The value of each part presented here
has been carefully chosen to avoid compromises with the working
of drone. As Table 1 succinctly summarizes, 3D printing of the
parts takes up the largest percentage of the overall cost. Hence, as
the cost of 3D printing comes down, so will the overall cost of the
system. The second most expensive item is observed to be the LiPo
battery. Other costly items in descending order are the motor and
the remote-control unit.
Table 1. Estimated cost of Manufacturing
Figure 16. Von-Mises results for 1.5 N
QUANTITY COST
COMPONENT
(UNITS) (US $)
3D Printing (Arms,
baseplate, motor mount, 14 272.47
drone arm)
Motors 4 72.22
Propeller 4 15.00
LiPo Battery 1 90.00
Electronic Speed controller
1 21.30
(ESC)
Camera 1 42.00
Arduino-Uno
1 24.99
Microcontroller
Remote control 1 49.99

Figure 17. Von-Mises results for 2.25 N Miscellaneous (Fasteners,

- 15.00
connectors)
Total $602.97

30th Florida Conference on Recent Advances in Robotics, May 11-12, 2017, Florida Atlantic University, Boca Raton, Florida
 6. DRONE WORKING ALGORITHM
Figure 18 represents the algorithm of the proposed drone. The
process of lifting the payload initially begins with switching on the
power supply from batteries to other electronic components such as
the motor and the electronic speed controller. As soon as the
systems and the microcontroller powers up the rotors, the speed is
adjusted for vertical take-off. The drone is now in air with the arm
preset to grab the required object.

The drone is designated with the position, which is monitored by

the controller (human) through the camera attached to it. Once the
location of the drone is set as per the instructions provided by the
controller, the drone’s speed is decreased and the end effector opens
and grabs the required object firmly with the help of suction cups
attached along the internal lining of the effector. This process is
repeated until the required object is obtained through series of trials
by the controller.

This problem can be eliminated in the future by the development of

autonomous self-guided system. The drone after grabbing the
object hovers towards the controller based on the instruction
specified by the operator. The drone is kept away at a safe distance
and the end effector based on designated position of the operator
drops the object in his/her vicinity. This process is repeated
continuously and the estimated endurance for the current proposed
drone model is about 15-20 min.

7. CONCLUSION
The function of the robot is defined in the simplest of way which
enables the UAV to perform a specific task. The materials and parts
selection have been considered based on detailed evaluation of
drones available in the market along with the mass of payload to be
carried. The end-effector is modeled such that it is equipped with
suction cups which give a better grip to the object help by the arm
also the arm is stationary with the links being adjusted manually.
The drone model presented here is controlled by means of a remote,
where further developments can be pursued to have a complete
autonomous system with capabilities of self-positioning, tracking
and voice control of the drone.

8. ACKNOWLEDGMENT
Our sincere thanks are extended to the Robotics and Automation
Laboratory in the Department of Mechanical and Materials
Engineering for providing access to the equipment and resources.

9. REFERENCES
1. Mendoza-Mendoza, Julio, et al. "Air-arm: A new kind of
flying manipulator." Research, Education and Development
of Unmanned Aerial Systems (RED-UAS), 2015 Workshop on.
IEEE, 2015.
2. Danko, Todd W., Kenneth P. Chaney, and Aul Y. Oh. "A
parallel manipulator for mobile manipulating
UAVs." Technologies for Practical Robot Applications
(TePRA), 2015 IEEE International Conference on. IEEE,
2015.
3. Danko, Todd W., and Paul Y. Oh. "A hyper-redundant
manipulator for mobile manipulating unmanned aerial
Figure 18. Flowchart for drone and end-effector control vehicles." Unmanned Aircraft Systems (ICUAS), 2013
International Conference on. IEEE, 2013.

30th Florida Conference on Recent Advances in Robotics, May 11-12, 2017, Florida Atlantic University, Boca Raton, Florida
4. Maier, Moritz, and Konstantin Kondak. "Landing of VTOL
UAVs using a stationary robot manipulator: An innovative
approach for coordinated control." Decision and Control
(CDC), 2015 IEEE 54th Annual Conference on. IEEE, 2015.
5. Zisimatos, Agisilaos G., et al. "Open-source, affordable,
modular, light-weight, underactuated robot hands." Intelligent
Robots and Systems (IROS 2014), 2014 IEEE/RSJ
International Conference on. IEEE, 2014.
6. T. Ueno, N. Sunaga, K. Brown and H. Asada., "Mechanism
and Control of a Dynamic Lifting Robot. " International
Association for Automation and Robotics in Construction
(ISARC 1996), 1996 ISARC International Conference.
30th Florida Conference on Recent Advances in Robotics, May 11-12, 2017, Florida Atlantic University, Boca Raton, Florida
7. Gund, Miss Shwetal Ramesh, and V. P. Bhope. "A Review on
Agricultural Robots." International Journal of Advanced
Research in Computer Engineering & Technology (IJARCET
2015).
8. Sarig, Y. Robotics of fruit harvesting: A state-of-the-art
review. Journal of agricultural engineering research, 54(4),
265-280, 1993.
9. Van Henten, E. J., Van Tuijl, B. V., Hemming, J., Kornet, J.
G., Bontsema, J., & Van Os, E. A. Field test of an autonomous
cucumber picking robot. Biosystems engineering, 86(3), 305-
313, 2003.  
10. Monkman, Gareth J., Stefan Hesse, Ralf Steinmann, and
Henrik Schunk. "Robot grippers". John Wiley & Sons, 2007.